Proof for Openclaw

Proof is a web-first collaborative markdown editor designed for seamless document interaction between AI agents and human users.

dshipper
v1.0.0
Mar 12, 2026
0
925
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install proofeditor

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install proofeditor using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Proof?

Proof serves as a specialized collaborative environment where agents and humans co-author markdown documents. By utilizing Openclaw Skills to connect to the Proof API, agents can join shared sessions, track provenance through authorship tags, and manage document states in real-time. This skill ensures that AI contributions are transparent and integrated into a professional workflow, moving beyond static text generation into active, multi-user document participation.

Unlike local file editing, Proof focuses on web-based collaboration using a centralized API at proofeditor.ai. It allows agents to show presence, provide suggestions via track changes, and manage complex document structures through block-level operations. This makes it an essential tool for teams using Openclaw Skills to build shared knowledge bases, reports, and iterative project documentation.

Proof Use Cases

  • Creating and sharing collaborative project plans, specs, or bug writeups.
  • Reviewing human-authored drafts and providing inline suggestions or comments.
  • Managing long-form documentation where provenance and edit history are required.
  • Coordinating multi-agent workflows where several AI entities contribute to a single source of truth.
  • Real-time brainstorming sessions where the agent maintains presence while the human edits.

How Proof Works

  1. The agent identifies a Proof URL or is triggered to create a new collaborative document.
  2. The skill extracts the unique document slug and access token to establish an authorized connection.
  3. The agent signals its presence using a specific agent ID and status (e.g., reading, thinking) to inform human collaborators.
  4. The agent fetches the current document state or a snapshot to understand the latest content and revision version.
  5. Edits are applied using various strategies, from simple appends to complex block-level operations with idempotency keys for reliability.
  6. The agent polls for pending events to react to human feedback, comments, or accepted suggestions.

Proof Setup

To use Proof with Openclaw Skills, ensure you have an API token from proofeditor.ai. You can create a new document using the following bash command:

curl -sS -X POST https://www.proofeditor.ai/share/markdown \
  -H "Content-Type: application/json" \
  -d '{"title":"My Document","markdown":"# Hello\n\nFirst draft."}'

For existing documents, use the Bearer token for authorization:

curl -sS "https://www.proofeditor.ai/api/agent/<slug>/state" \
  -H "Authorization: Bearer <token>" \
  -H "X-Agent-Id: <your-agent-id>"

Proof Data Schema & Taxonomy

Proof organizes data through document slugs and snapshots. Key metadata elements include:

Field Description
slug The unique identifier for the document URL
accessToken The credential required for API interactions
by Attribution string (e.g., ai:agent-name) required for every write
X-Agent-Id Used to manage presence identity and status
baseRevision Version number used to prevent stale overwrites (409 errors)

Supported block operations include replace_block, insert_before, insert_after, delete_block, and find_replace_in_block.

Proof Advanced Features

  • Idempotent Edits: Uses Idempotency-Key in V2 edits to ensure safe retries without duplicating content.
  • Track Changes: Supports suggestion.add and rewrite.apply for human-in-the-loop review cycles.
  • Real-time Presence: Agents can update their status to reading, thinking, acting, or waiting.
  • Revision Locking: Prevents data loss by checking the baseRevision against the current server state.
  • Event Acknowledgment: Allows agents to process and ack specific document events to stay in sync.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*